Speech synthesis is the production of speech from text by artificial means. For example, text-to-speech (TTS) systems synthesize speech from text to provide an alternative to conventional computer-to-human visual output devices like computer monitors or displays. One problem encountered with TTS synthesis is that synthesized speech can have poor prosodic characteristics, such as intonation, pronunciation, stress, speaking rate, tone, and naturalness. Accordingly, such poor prosody can confuse a TTS user and result in incomplete interaction with the user.